This work aimed to determinate an adequate methodology to evaluate plant abundance in an area of Ipomoea pes-caprae, taking Navegantes dunes (SC, 26o54’40"S e 48o39’00"W) as a case study. Abundance as percentage of coverage was determinate using a sampler divided by nylon filaments. A stick was put at the filaments intersections and the species touching the stick was recorded. We tested 3 sampler sizes: 1m2, 0.56m2 and 0.25m2. For each sampler two different samplings were made: the species were recorded in all intersections and also in 30 randomly selected intersections. A dune zone 50mX10m was delimited, where 15 replicates for each treatment were sampled. We calculated average of coverage and standard error as percentage, for 2 replicates, 3, 4, until 15 replicates, for each treatment. The use of all intersects or 30 intersections presented no influences on the results. Coverage percentages for the two smaller samplers were about 30%, but standard error was higher to 0.25m2 sampler (about 20%), then to others (about 10%). Coverage percentage obtained by 1m2 sampler was overestimated (40%) due to difficulties to take a vertical look at the intersections far from the sampler edge. Six to eight replicates were needed when using the 0.56m2 and 1m2 samplers, and ten replicates were needed when using 0.25m2 sampler. Considering these results, we concluded the best sampling is made using 0.56m2 sampler, 30 intersections randomly selected and 6 to 8 replicates. These make possible to optimize the data collection with the minimum expenses of time and effort, and also the best description of the vegetation cover.